Nothing can contain how much arghage is in this thread

EnverexEnverex Worcester, UK Icrontian
edited October 2003 in Folding@Home
Damnit, because F@H was not executed from its own directory on my Linux machine (executed from current directory at startup (being root) rather than navigating to that directory first then executing it) it has been folding under Anon. I was wondering why my points were still down even after getting that machine back on.

Even though it's an AXP2000 it is still going to be slow though until Stanford sort themselves out and make a working 3.25 client as the current one doesnt work on most Linux machines, thus no SSE boost.

:banghead:

NS

EDIT: Edited for extreme unreadableness

Comments

  • ShortyShorty Manchester, UK Icrontian
    edited October 2003
    Sucks x 2 :(
  • edcentricedcentric near Milwaukee, Wisconsin Icrontian
    edited October 2003
    argh argh
  • Park_7677Park_7677 Missouri Member
    edited October 2003
    I've done this 100 times :doh:

    GL gettin the points back up there.
  • EnverexEnverex Worcester, UK Icrontian
    edited October 2003
    On that note, is anyone here running F@H through Wine in Linux as apparently Linux folds 30% slower than Windows, but folds at the same speed if you run it through one of the emulators.

    It may be worth starting a new thread for this instead though.

    NS
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited October 2003
    Well, mine can use (and does) SSE and it is on a P4 with the 3.24 Large RAM Beta client and an Enterprise kernel. Problem is unless you have an SMP or Enterprise kernel you will not take advantage of any SMP or multipipe features. Yes, mine runs from /home/john/FAH2. But it DOES run SSE and does run the tings with the advmethods switch also, what it does NOT like is the -forceasm (force advanced assembly code) options as simply those do not work in Linux on my P4 and give me a 20-30% DECREASE in folding throughput if used on my P4. Since Linux likes to determine assembly options used, this is not surprising. So, I fold only with the -advmethods switch on my P4 box.

    The 3.25 client simply segfaults.

    John.
  • EnverexEnverex Worcester, UK Icrontian
    edited October 2003
    Ageek said
    Well, mine can use (and does) SSE and it is on a P4 with the 3.24 Large RAM Beta client and an Enterprise kernel. Problem is unless you have an SMP or Enterprise kernel you will not take advantage of any SMP or multipipe features. Yes, mine runs from /home/john/FAH2. But it DOES run SSE and does run the tings with the advmethods switch also, what it does NOT like is the -forceasm (force advanced assembly code) options as simply those do not work in Linux on my P4 and give me a 20-30% DECREASE in folding throughput if used on my P4. Since Linux likes to determine assembly options used, this is not surprising. So, I fold only with the -advmethods switch on my P4 box.

    The 3.25 client simply segfaults.

    John.

    You dont NEED to use the 3.25 client because P4's automatically use SSE without requiring the -forceasm swtich, and when that switch is used, it normally causes the client to crash (on P4 machines). So thats why you dont need to use it.

    The switches work perfectly in Linux, they just dont work with your processor. Check the other threads regarding the switches and what to use with each processor.

    The segfault error is what I suffer from with 3.25 and that is the reason I cant use it. This can be solved by using an older GCC compliler, but I do not wish to downgrade.

    I fail to see what enterprise kernels, SMP and piping have to do with anything in this thread though.

    NS
  • mmonninmmonnin Centreville, VA
    edited October 2003
    Never used Linux before so I am no help. And Ageek, his is a AMD not a P4 like yours.
Sign In or Register to comment.